Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

switch to row_major default #195

Open
wants to merge 1 commit into
base: master
Choose a base branch
from
Open

switch to row_major default #195

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
22 changes: 20 additions & 2 deletions docs/source/array_tensor.rst
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -4,8 +4,8 @@

The full license is in the file LICENSE, distributed with this software.

Arrays and tensors
==================
Arrays, Tensors and NumPy Views
===============================

``xtensor-python`` provides two container types wrapping numpy arrays: ``pyarray`` and ``pytensor``. They are the counterparts
to ``xarray`` and ``xtensor`` containers.
Expand All @@ -25,3 +25,21 @@ Like ``xtensor``, ``pytensor`` has a static stack-allocated shape. This means th
the shape of the ``pytensor`` upon creation. As a consequence, reshapes are not reflected across languages. However, this drawback
is offset by a more effective computation of shape and broadcast.

NumPy views
-----------

If you are trying to call a xtensor function with a NumPy view, you might read the following:
"passing container with wrong strides for static layout". This means, that the *static layout*
of the pytensor does not match the layout of the NumPy array (or view) that you are using to
call the function.

One fix for this is to choose the ``dynamic`` layout for your pyarray or pytensor:

.. code::

void my_function(xt::pyarray<double, xt::layout_type::dynamic> input)

When choosing the dynamic layout, you can pass any NumPy container (with Fortran order, or
an arbitrary view).
Note that iterators on a pyarray with dynamic layout are generally slower!

8 changes: 4 additions & 4 deletions include/xtensor-python/pyarray.hpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -23,7 +23,7 @@

namespace xt
{
template <class T, layout_type L = layout_type::dynamic>
template <class T, layout_type L = layout_type::row_major>
class pyarray;
}

Expand Down Expand Up @@ -682,7 +682,7 @@ namespace xt

if (!tmp)
{
throw std::runtime_error("NumPy: unable to create ndarray");
throw std::runtime_error("pyarray: unable to create ndarray");
}

this->m_ptr = tmp.release().ptr();
Expand Down Expand Up @@ -771,7 +771,7 @@ namespace xt

if (!tmp)
{
throw std::runtime_error("NumPy: unable to create ndarray");
throw std::runtime_error("pyarray: unable to create ndarray");
}

this->m_ptr = tmp.release().ptr();
Expand All @@ -793,7 +793,7 @@ namespace xt

if (L != layout_type::dynamic && !do_strides_match(m_shape, m_strides, L, 1))
{
throw std::runtime_error("NumPy: passing container with bad strides for layout (is it a view?).");
throw std::runtime_error("pyarray: passing container with wrong strides for static layout. If you are trying to pass a view, set the pytensor layout template parameter to `layout_type::dynamic`, or `layout_type::column_major` for col major.");
}

m_backstrides = backstrides_type(*this);
Expand Down
8 changes: 4 additions & 4 deletions include/xtensor-python/pytensor.hpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -24,7 +24,7 @@

namespace xt
{
template <class T, std::size_t N, layout_type L = layout_type::dynamic>
template <class T, std::size_t N, layout_type L = layout_type::row_major>
class pytensor;
}

Expand Down Expand Up @@ -431,7 +431,7 @@ namespace xt

if (!tmp)
{
throw std::runtime_error("NumPy: unable to create ndarray");
throw std::runtime_error("pytensor: unable to create ndarray");
}

this->m_ptr = tmp.release().ptr();
Expand All @@ -452,7 +452,7 @@ namespace xt

if (PyArray_NDIM(this->python_array()) != N)
{
throw std::runtime_error("NumPy: ndarray has incorrect number of dimensions");
throw std::runtime_error("pytensor: ndarray has incorrect number of dimensions");
}

std::copy(PyArray_DIMS(this->python_array()), PyArray_DIMS(this->python_array()) + N, m_shape.begin());
Expand All @@ -462,7 +462,7 @@ namespace xt

if (L != layout_type::dynamic && !do_strides_match(m_shape, m_strides, L, 1))
{
throw std::runtime_error("NumPy: passing container with bad strides for layout (is it a view?).");
throw std::runtime_error("pytensor: passing container with wrong strides for static layout. If you are trying to pass a view, set the pytensor layout template parameter to `layout_type::dynamic`, or `layout_type::column_major` for col major.");
}

m_storage = storage_type(reinterpret_cast<pointer>(PyArray_DATA(this->python_array())),
Expand Down